// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
// REQUIRES: amdgpu-registered-target
// RUN: %clang_cc1 %s -x hip -fcuda-is-device -emit-llvm -O0 -o - \
// RUN:   -triple=amdgcn-amd-amdhsa  | opt -S | FileCheck %s

// CHECK-LABEL: @_Z29test_non_volatile_parameter32Pj(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[PTR_ADDR:%.*]] = alloca i32*, align 8, addrspace(5)
// CHECK-NEXT:    [[RES:%.*]] = alloca i32, align 4, addrspace(5)
// CHECK-NEXT:    [[PTR_ADDR_ASCAST:%.*]] = addrspacecast i32* addrspace(5)* [[PTR_ADDR]] to i32**
// CHECK-NEXT:    [[RES_ASCAST:%.*]] = addrspacecast i32 addrspace(5)* [[RES]] to i32*
// CHECK-NEXT:    store i32* [[PTR:%.*]], i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* [[TMP1]], align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* [[TMP0]], i32 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP3]], i32* [[RES_ASCAST]], align 4
// CHECK-NEXT:    [[TMP4:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
// CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* [[TMP4]], i32 [[TMP6]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP7]], i32* [[RES_ASCAST]], align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_non_volatile_parameter32(__UINT32_TYPE__ *ptr) {
  __UINT32_TYPE__ res;
  res = __builtin_amdgcn_atomic_inc32(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");

  res = __builtin_amdgcn_atomic_dec32(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z29test_non_volatile_parameter64Py(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[PTR_ADDR:%.*]] = alloca i64*, align 8, addrspace(5)
// CHECK-NEXT:    [[RES:%.*]] = alloca i64, align 8, addrspace(5)
// CHECK-NEXT:    [[PTR_ADDR_ASCAST:%.*]] = addrspacecast i64* addrspace(5)* [[PTR_ADDR]] to i64**
// CHECK-NEXT:    [[RES_ASCAST:%.*]] = addrspacecast i64 addrspace(5)* [[RES]] to i64*
// CHECK-NEXT:    store i64* [[PTR:%.*]], i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP0:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP1:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i64, i64* [[TMP1]], align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* [[TMP0]], i64 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP3]], i64* [[RES_ASCAST]], align 8
// CHECK-NEXT:    [[TMP4:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP5:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP6:%.*]] = load i64, i64* [[TMP5]], align 8
// CHECK-NEXT:    [[TMP7:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* [[TMP4]], i64 [[TMP6]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP7]], i64* [[RES_ASCAST]], align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_non_volatile_parameter64(__UINT64_TYPE__ *ptr) {
  __UINT64_TYPE__ res;
  res = __builtin_amdgcn_atomic_inc64(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");

  res = __builtin_amdgcn_atomic_dec64(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z25test_volatile_parameter32PVj(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[PTR_ADDR:%.*]] = alloca i32*, align 8, addrspace(5)
// CHECK-NEXT:    [[RES:%.*]] = alloca i32, align 4, addrspace(5)
// CHECK-NEXT:    [[PTR_ADDR_ASCAST:%.*]] = addrspacecast i32* addrspace(5)* [[PTR_ADDR]] to i32**
// CHECK-NEXT:    [[RES_ASCAST:%.*]] = addrspacecast i32 addrspace(5)* [[RES]] to i32*
// CHECK-NEXT:    store i32* [[PTR:%.*]], i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP1:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load volatile i32, i32* [[TMP1]], align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* [[TMP0]], i32 [[TMP2]], i32 7, i32 2, i1 true)
// CHECK-NEXT:    store i32 [[TMP3]], i32* [[RES_ASCAST]], align 4
// CHECK-NEXT:    [[TMP4:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP6:%.*]] = load volatile i32, i32* [[TMP5]], align 4
// CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* [[TMP4]], i32 [[TMP6]], i32 7, i32 2, i1 true)
// CHECK-NEXT:    store i32 [[TMP7]], i32* [[RES_ASCAST]], align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_volatile_parameter32(volatile __UINT32_TYPE__ *ptr) {
  __UINT32_TYPE__ res;
  res = __builtin_amdgcn_atomic_inc32(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");

  res = __builtin_amdgcn_atomic_dec32(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z25test_volatile_parameter64PVy(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[PTR_ADDR:%.*]] = alloca i64*, align 8, addrspace(5)
// CHECK-NEXT:    [[RES:%.*]] = alloca i64, align 8, addrspace(5)
// CHECK-NEXT:    [[PTR_ADDR_ASCAST:%.*]] = addrspacecast i64* addrspace(5)* [[PTR_ADDR]] to i64**
// CHECK-NEXT:    [[RES_ASCAST:%.*]] = addrspacecast i64 addrspace(5)* [[RES]] to i64*
// CHECK-NEXT:    store i64* [[PTR:%.*]], i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP0:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP1:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load volatile i64, i64* [[TMP1]], align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* [[TMP0]], i64 [[TMP2]], i32 7, i32 2, i1 true)
// CHECK-NEXT:    store i64 [[TMP3]], i64* [[RES_ASCAST]], align 8
// CHECK-NEXT:    [[TMP4:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP5:%.*]] = load i64*, i64** [[PTR_ADDR_ASCAST]], align 8
// CHECK-NEXT:    [[TMP6:%.*]] = load volatile i64, i64* [[TMP5]], align 8
// CHECK-NEXT:    [[TMP7:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* [[TMP4]], i64 [[TMP6]], i32 7, i32 2, i1 true)
// CHECK-NEXT:    store i64 [[TMP7]], i64* [[RES_ASCAST]], align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_volatile_parameter64(volatile __UINT64_TYPE__ *ptr) {
  __UINT64_TYPE__ res;
  res = __builtin_amdgcn_atomic_inc64(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");

  res = __builtin_amdgcn_atomic_dec64(ptr, *ptr, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z13test_shared32v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ13test_shared32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ13test_shared32vE3val to i32*), i32 [[TMP0]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP1]], i32* addrspacecast (i32 addrspace(3)* @_ZZ13test_shared32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ13test_shared32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ13test_shared32vE3val to i32*), i32 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP3]], i32* addrspacecast (i32 addrspace(3)* @_ZZ13test_shared32vE3val to i32*), align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_shared32() {
  __attribute__((shared)) __UINT32_TYPE__ val;

  val = __builtin_amdgcn_atomic_inc32(&val, val, __ATOMIC_SEQ_CST, "workgroup");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z13test_shared64v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ13test_shared64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ13test_shared64vE3val to i64*), i64 [[TMP0]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP1]], i64* addrspacecast (i64 addrspace(3)* @_ZZ13test_shared64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ13test_shared64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ13test_shared64vE3val to i64*), i64 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP3]], i64* addrspacecast (i64 addrspace(3)* @_ZZ13test_shared64vE3val to i64*), align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_shared64() {
  __attribute__((shared)) __UINT64_TYPE__ val;

  val = __builtin_amdgcn_atomic_inc64(&val, val, __ATOMIC_SEQ_CST, "workgroup");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_SEQ_CST, "workgroup");
}

__attribute__((device)) __UINT32_TYPE__ global_val32;
// CHECK-LABEL: @_Z13test_global32v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i32, i32* addrspacecast (i32 addrspace(1)* @global_val32 to i32*), align 4
// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(1)* @global_val32 to i32*), i32 [[TMP0]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP1]], i32* addrspacecast (i32 addrspace(1)* @global_val32 to i32*), align 4
// CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* addrspacecast (i32 addrspace(1)* @global_val32 to i32*), align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(1)* @global_val32 to i32*), i32 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP3]], i32* addrspacecast (i32 addrspace(1)* @global_val32 to i32*), align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_global32() {
  global_val32 = __builtin_amdgcn_atomic_inc32(&global_val32, global_val32, __ATOMIC_SEQ_CST, "workgroup");

  global_val32 = __builtin_amdgcn_atomic_dec32(&global_val32, global_val32, __ATOMIC_SEQ_CST, "workgroup");
}

__attribute__((device)) __UINT64_TYPE__ global_val64;
// CHECK-LABEL: @_Z13test_global64v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i64, i64* addrspacecast (i64 addrspace(1)* @global_val64 to i64*), align 8
// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* addrspacecast (i64 addrspace(1)* @global_val64 to i64*), i64 [[TMP0]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP1]], i64* addrspacecast (i64 addrspace(1)* @global_val64 to i64*), align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i64, i64* addrspacecast (i64 addrspace(1)* @global_val64 to i64*), align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(1)* @global_val64 to i64*), i64 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP3]], i64* addrspacecast (i64 addrspace(1)* @global_val64 to i64*), align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_global64() {
  global_val64 = __builtin_amdgcn_atomic_inc64(&global_val64, global_val64, __ATOMIC_SEQ_CST, "workgroup");

  global_val64 = __builtin_amdgcn_atomic_dec64(&global_val64, global_val64, __ATOMIC_SEQ_CST, "workgroup");
}

__attribute__((constant)) __UINT32_TYPE__ cval32;
// CHECK-LABEL: @_Z15test_constant32v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[LOCAL_VAL:%.*]] = alloca i32, align 4, addrspace(5)
// CHECK-NEXT:    [[LOCAL_VAL_ASCAST:%.*]] = addrspacecast i32 addrspace(5)* [[LOCAL_VAL]] to i32*
// CHECK-NEXT:    [[TMP0:%.*]] = load i32, i32* addrspacecast (i32 addrspace(4)* @cval32 to i32*), align 4
// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(4)* @cval32 to i32*), i32 [[TMP0]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP1]], i32* [[LOCAL_VAL_ASCAST]], align 4
// CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* addrspacecast (i32 addrspace(4)* @cval32 to i32*), align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(4)* @cval32 to i32*), i32 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP3]], i32* [[LOCAL_VAL_ASCAST]], align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_constant32() {
  __UINT32_TYPE__ local_val;

  local_val = __builtin_amdgcn_atomic_inc32(&cval32, cval32, __ATOMIC_SEQ_CST, "workgroup");

  local_val = __builtin_amdgcn_atomic_dec32(&cval32, cval32, __ATOMIC_SEQ_CST, "workgroup");
}

__attribute__((constant)) __UINT64_TYPE__ cval64;
// CHECK-LABEL: @_Z15test_constant64v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[LOCAL_VAL:%.*]] = alloca i64, align 8, addrspace(5)
// CHECK-NEXT:    [[LOCAL_VAL_ASCAST:%.*]] = addrspacecast i64 addrspace(5)* [[LOCAL_VAL]] to i64*
// CHECK-NEXT:    [[TMP0:%.*]] = load i64, i64* addrspacecast (i64 addrspace(4)* @cval64 to i64*), align 8
// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* addrspacecast (i64 addrspace(4)* @cval64 to i64*), i64 [[TMP0]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP1]], i64* [[LOCAL_VAL_ASCAST]], align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i64, i64* addrspacecast (i64 addrspace(4)* @cval64 to i64*), align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(4)* @cval64 to i64*), i64 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP3]], i64* [[LOCAL_VAL_ASCAST]], align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_constant64() {
  __UINT64_TYPE__ local_val;

  local_val = __builtin_amdgcn_atomic_inc64(&cval64, cval64, __ATOMIC_SEQ_CST, "workgroup");

  local_val = __builtin_amdgcn_atomic_dec64(&cval64, cval64, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z12test_order32v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), i32 [[TMP0]], i32 2, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP1]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), i32 [[TMP2]], i32 4, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP3]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP4:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP5:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), i32 [[TMP4]], i32 4, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP5]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP6:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), i32 [[TMP6]], i32 5, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP7]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP8:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP9:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), i32 [[TMP8]], i32 6, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP9]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP10:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP11:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), i32 [[TMP10]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP11]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_order32vE3val to i32*), align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_order32() {
  __attribute__((shared)) __UINT32_TYPE__ val;

  val = __builtin_amdgcn_atomic_inc32(&val, val, __ATOMIC_RELAXED, "workgroup");

  val = __builtin_amdgcn_atomic_inc32(&val, val, __ATOMIC_CONSUME, "workgroup");

  val = __builtin_amdgcn_atomic_inc32(&val, val, __ATOMIC_ACQUIRE, "workgroup");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_RELEASE, "workgroup");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_ACQ_REL, "workgroup");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z12test_order64v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), i64 [[TMP0]], i32 2, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP1]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), i64 [[TMP2]], i32 4, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP3]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP4:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP5:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), i64 [[TMP4]], i32 4, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP5]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP6:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP7:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), i64 [[TMP6]], i32 5, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP7]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP8:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP9:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), i64 [[TMP8]], i32 6, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP9]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP10:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP11:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), i64 [[TMP10]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP11]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_order64vE3val to i64*), align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_order64() {
  __attribute__((shared)) __UINT64_TYPE__ val;

  val = __builtin_amdgcn_atomic_inc64(&val, val, __ATOMIC_RELAXED, "workgroup");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_CONSUME, "workgroup");

  val = __builtin_amdgcn_atomic_inc64(&val, val, __ATOMIC_ACQUIRE, "workgroup");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_RELEASE, "workgroup");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_ACQ_REL, "workgroup");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_SEQ_CST, "workgroup");
}

// CHECK-LABEL: @_Z12test_scope32v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP1:%.*]] = call i32 @llvm.amdgcn.atomic.inc.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), i32 [[TMP0]], i32 7, i32 1, i1 false)
// CHECK-NEXT:    store i32 [[TMP1]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP3:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), i32 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i32 [[TMP3]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP4:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP5:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), i32 [[TMP4]], i32 7, i32 3, i1 false)
// CHECK-NEXT:    store i32 [[TMP5]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP6:%.*]] = load i32, i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    [[TMP7:%.*]] = call i32 @llvm.amdgcn.atomic.dec.i32.p0i32(i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), i32 [[TMP6]], i32 7, i32 4, i1 false)
// CHECK-NEXT:    store i32 [[TMP7]], i32* addrspacecast (i32 addrspace(3)* @_ZZ12test_scope32vE3val to i32*), align 4
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_scope32() {
  __attribute__((shared)) __UINT32_TYPE__ val;

  val = __builtin_amdgcn_atomic_inc32(&val, val, __ATOMIC_SEQ_CST, "");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_SEQ_CST, "workgroup");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_SEQ_CST, "agent");

  val = __builtin_amdgcn_atomic_dec32(&val, val, __ATOMIC_SEQ_CST, "wavefront");
}

// CHECK-LABEL: @_Z12test_scope64v(
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[TMP0:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP1:%.*]] = call i64 @llvm.amdgcn.atomic.inc.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), i64 [[TMP0]], i32 7, i32 1, i1 false)
// CHECK-NEXT:    store i64 [[TMP1]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP2:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), i64 [[TMP2]], i32 7, i32 2, i1 false)
// CHECK-NEXT:    store i64 [[TMP3]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP4:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP5:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), i64 [[TMP4]], i32 7, i32 3, i1 false)
// CHECK-NEXT:    store i64 [[TMP5]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP6:%.*]] = load i64, i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    [[TMP7:%.*]] = call i64 @llvm.amdgcn.atomic.dec.i64.p0i64(i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), i64 [[TMP6]], i32 7, i32 4, i1 false)
// CHECK-NEXT:    store i64 [[TMP7]], i64* addrspacecast (i64 addrspace(3)* @_ZZ12test_scope64vE3val to i64*), align 8
// CHECK-NEXT:    ret void
//
__attribute__((device)) void test_scope64() {
  __attribute__((shared)) __UINT64_TYPE__ val;

  val = __builtin_amdgcn_atomic_inc64(&val, val, __ATOMIC_SEQ_CST, "");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_SEQ_CST, "workgroup");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_SEQ_CST, "agent");

  val = __builtin_amdgcn_atomic_dec64(&val, val, __ATOMIC_SEQ_CST, "wavefront");
}